The Human Cytoplasmic Dynein Interactome Reveals Novel Activators
Total Page:16
File Type:pdf, Size:1020Kb
TOOLS AND RESOURCES The human cytoplasmic dynein interactome reveals novel activators of motility William B Redwine1,2†, Morgan E DeSantis1†, Ian Hollyer1‡, Zaw Min Htet1,3, Phuoc Tien Tran1, Selene K Swanson4, Laurence Florens4, Michael P Washburn4,5, Samara L Reck-Peterson1,6* 1Department of Cellular and Molecular Medicine, University of California, San Diego, United States; 2Department of Cell Biology, Harvard Medical School, Boston, United States; 3Biophysics Graduate Program, Harvard Medical School, Boston, United States; 4Stowers Institute for Medical Research, Kansas, United States; 5Department of Pathology and Laboratory Medicine, The University of Kansas Medical Center, Kansas, United States; 6Division of Biological Sciences, Cell and Developmental Biology Section, University of California, San Diego, United States Abstract In human cells, cytoplasmic dynein-1 is essential for long-distance transport of many cargos, including organelles, RNAs, proteins, and viruses, towards microtubule minus ends. To understand how a single motor achieves cargo specificity, we identified the human dynein *For correspondence: interactome by attaching a promiscuous biotin ligase (‘BioID’) to seven components of the dynein [email protected] machinery, including a subunit of the essential cofactor dynactin. This method reported spatial †These authors contributed information about the large cytosolic dynein/dynactin complex in living cells. To achieve maximal equally to this work motile activity and to bind its cargos, human dynein/dynactin requires ‘activators’, of which only five have been described. We developed methods to identify new activators in our BioID data, and Present address: ‡Feinberg discovered that ninein and ninein-like are a new family of dynein activators. Analysis of the protein School of Medicine, Northwestern University, interactomes for six activators, including ninein and ninein-like, suggests that each dynein activator Chicago, United States has multiple cargos. DOI: 10.7554/eLife.28257.001 Competing interest: See page 22 Funding: See page 23 Introduction Received: 01 May 2017 Microtubules and their motors are the primary means of long-distance intracellular transport in Accepted: 14 July 2017 humans and many other eukaryotic organisms. Mutations in the transport machinery cause both neu- Published: 18 July 2017 rodevelopmental and neurodegenerative diseases (Lipka et al., 2013). Microtubules are polar struc- Reviewing editor: Mohan K tures, with dynamic ‘plus’ ends typically found near the cell periphery and ‘minus’ ends anchored in Balasubramanian, University of internal microtubule organizing centers. Dynein motors move towards the microtubule minus end, Warwick, United Kingdom whereas most kinesins move in the opposite direction. The human genome contains 15 motor domain-encoding dynein genes (Vale, 2003), but only cytoplasmic dynein-1 (DYNC1H1; ‘dynein’ Copyright Redwine et al. This hereafter) is involved in long-distance, minus-end-directed transport in the cytoplasm. Dynein trans- article is distributed under the ports dozens of distinct cargos including organelles, ribonucleoprotein complexes, proteins and terms of the Creative Commons Attribution License, which viruses (Kardon and Vale, 2009). A major outstanding question in the field is to understand how permits unrestricted use and dynein achieves temporal and spatial specificity for cargo interactions. redistribution provided that the Most cytoskeletal motors that transport cargos over long distances in cells are processive motors, original author and source are capable of taking multiple steps along their track. While dimers of the S. cerevisiae dynein heavy credited. chain move processively in the absence of cofactors (Reck-Peterson et al., 2006), mammalian Redwine et al. eLife 2017;6:e28257. DOI: 10.7554/eLife.28257 1 of 27 Tools and resources Cell Biology dynein requires the 1.1 MDa dynactin complex and a coiled coil-containing activator (‘activator’ hereafter) for robust processive motility (McKenney et al., 2014; Schlager et al., 2014; Trokter et al., 2012). Activators have a second function; they also link dynein/dynactin to cargo (Figure 1A)(Cianfrocco et al., 2015). Currently, there are five proteins that likely function as dynein activators. The activators BICD2 and HOOK3 have been definitively shown, using purified components, to activate dynein/dynactin motility in vitro (McKenney et al., 2014; Schlager et al., 2014; Schroeder and Vale, 2016). HOOK1, Spindly (SPDL1), and RAB11FIP3 are also likely activators based on their ability to co-purify and co-migrate with dynein/dynactin in in vitro motility assays (McKenney et al., 2014; Olenick et al., 2016). Other proteins may be activators based on their homology to BICD and HOOK family activators, including BICD1, BICDL1, BICDL2, and HOOK2 (Hoogenraad and Akhma- nova, 2016; Simpson et al., 2005). Known and predicted activators all contain long stretches of pre- dicted coiled coil, but share very little sequence homology across activator families (Gama et al., 2017); currently it is not possible to identify activators based on sequence alone. Central to under- standing how dynein performs so many tasks is to determine if it has additional activators. Here we used new proteomics tools to address major unanswered questions about dynein-based transport. What is the dynein protein interactome? How many activators does dynein have in a given cell type? Which cargos do activators link to? Does each cargo have its own activator? To answer these questions, we used proximity-dependent labeling in living human cells. Traditionally, protein- protein interaction discovery using immunoprecipitation followed by mass spectrometry has been confined to relatively stable interactions. However, recently developed methods such as BioID (Roux et al., 2012) and APEX (Rhee et al., 2013) allow the discovery of weak and short-lived inter- actions in living cells, in addition to more stable interactions. The BioID method relies on expressing a protein of interest fused to a promiscuous biotin ligase that releases activated biotin-AMP in the absence of substrate (Roux et al., 2012). Biotin-AMP covalently modifies the primary amines of proximal proteins within a nanometer-scale labeling radius (Kim et al., 2014). Biotinlyated proximal proteins are identified by isolation with streptavidin followed by tandem mass spectrometry (MS/ MS). For example, this approach has been used to map protein interactions at human centrosomes and cilia (Gupta et al., 2015), focal adhesions (Dong et al., 2016) and the nuclear pore (Kim et al., 2014). Using these methods we describe the human dynein/dynactin interactome. We also developed methods to identify dynein activators within these datasets and identified two new activators that constitute a novel activator family. Finally, to determine the candidate cargos of six distinct activa- tors we elucidated their individual interactomes. Our results suggest that each dynein activator has multiple cargos. We propose that activators provide the first layer of defining cargo specificity for cytoplasmic dynein, but that refinement of cargo selection will require additional factors. Results Identification of the dynein/dynactin interactome To identify the human dynein/dynactin interactome, we began by biochemically characterizing dynein and dynactin subunits fused to BioID that were stably expressed in HEK-293 cells. The 1.4 MDa dynein holoenzyme is composed of dimers of heavy chains (HC; DYNC1H1), intermediate chains (IC1 or IC2; DYNC1I1 and 2), light intermediate chains (LIC1 or LIC2; DYNC1LI1 and 2), and three types of light chains: Roadblock (RB; DYNLRB1 and 2), LC8 (DYNLL1 and 2), and TCTEX (DYNLT1 and 3) (Figure 1A and Figure 1—figure supplement 1). We first generated a cell line sta- bly expressing IC2 with C-terminal BioID G2 (‘BioID’ here) (Kim et al., 2016b) and 3ÂFLAG tags. Immunoprecipitations confirmed that IC2-BioID was incorporated into the dynein/dynactin complex (Figure 1B—E). Gel filtration analysis of IC2 immunoprecipitates revealed that 51% of the BioID- tagged IC2 was incorporated into the dynein complex (Figure 1E). We obtained similar results when BioID was fused to the C-terminus of the p62 dynactin subunit. The stably expressed p62-BioID- 3ÂFLAG subunit incorporated into the dynactin complex as shown by immunoprecipitations (Figure 1F), and gel filtration analysis of these immunoprecipitatates revealed that 47% was incorpo- rated into the high molecular weight dynactin complex (Figure 1G). Redwine et al. eLife 2017;6:e28257. DOI: 10.7554/eLife.28257 2 of 27 Tools and resources Cell Biology Figure 1. Validation of BioID-tagged dynein and dynactin subunits. (A) A cartoon of the dynein/dynactin/activator complex based on cryo-EM structural studies (Chowdhury et al., 2015; Urnavicius et al., 2015) with proteins drawn to scale. (B) BioID-3ÂFLAG or IC2-BioID-3ÂFLAG were immunoprecipitated from stable HEK-293 cell lines using a-FLAG antibodies and eluted using FLAG peptide. A Sypro Red stained SDS-PAGE gel of the immunoprecipitates is shown. (C) MS/MS analysis of the immunoprecipitates from (B). Core dynein subunit dNSAF (distributed normalized spectral abundance factor) (Zhang et al., 2015) values are displayed as a gray scale heat map. (D) Immunoprecipitations were performed as in (B) with mild (M) or harsh (H) detergent conditions (see Figure 1 continued on next page Redwine et al. eLife 2017;6:e28257. DOI: 10.7554/eLife.28257 3 of 27 Tools and resources Cell Biology